Back to skill

Security audit

Speak AI

Security checks for vulnerabilities and agentic risk

Overview

The Speak AI workflow is sensible, but setup can run downloaded software and introduce additional agent skills without a clear approval step.

Review setup before enabling it. Prefer preinstalling a verified CLI release and approving its local changes explicitly. To prevent the documented startup skill synchronization, consider setting OO_SKILLS_SYNC_DISABLED=1 before invoking oo. OOMOL mediates access to your connected Speak AI account. The remote installers were unavailable for inspection, so their exact effects remain unverified.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared purpose concerns Speak AI data operations, but the supplied code executes an installer from cli.oomol.com. Nothing in the chunk establishes that this installation supports Speak AI, so it represents an undeclared capability and an inconsistent resource. The remote script's contents are not supplied, so its further behavior cannot be assessed.

Content

No source excerpt is available for this finding.

External Script Fetching

High
Category
Supply Chain
Confidence
97% confidence
Finding

Piping a remotely fetched script directly into bash executes mutable code without an independent integrity check or review; the PowerShell alternative has the same weakness. Although the domain is relevant and HTTPS protects transport, compromise of the installer distribution could yield arbitrary code execution with the invoking user's permissions.

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

Fallback setup executes remote installer scripts without warning about local system changes.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The fallback instructs the agent to install software after a missing-command error without first disclosing local changes or obtaining approval; the confirmation requirement elsewhere applies only to Speak AI write actions. The error-triggered condition limits exposure, but installing a dependency is a separate host-level action that can modify files and configuration.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 81)May include surrounding context.

md
- **`OOMOL_INSUFFICIENT_CREDIT`** — OOMOL billing stop. Recharge at `https://console.oomol.com/billing/token-recharge` before retrying.

- **HTTP 402 without `OOMOL_INSUFFICIENT_CREDIT`** — the third-party account itself has no active plan or credit. Report the provider error message as is; do not ask the user to recharge OOMOL.

## Resources